# Problems and Solutions: CRMO-2012, Paper 4

1. Let ABCD be a unit square. Draw a quadrant of a circle with A as centre and B, D as end points of the arc. Similarly, draw a quadrant of a circle with B as centre and A, C as end points of the arc. Inscribe a circle Γ touching the arc AC externally, the arc BD internally and also touching the side AD. Find the radius of the circle Γ. Solution: Let O be the centre of Γ and r its radius. Draw OP ⊥ AD and OQ ⊥ AB . Then OP = r, OQ2 = OA2 − r2 = (1 − r)2 − r2 = 1 − 2r. We also have OB = 1 + r and BQ = 1 − r. Using Pythagoras’ theorem we get (1 + r)2 = (1 − r)2 + 1 − 2r. Simpliﬁcation gives r = 1/6. 2. Let a, b, c be positive integers such that a divides b2 , b divides c2 and c divides a2 . Prove that abc divides (a + b + c)7 . Solution: If a prime p divides a, then p | b2 and hence p | b. This implies that p | c2 and hence p | c. Thus every prime dividing a also divides b and c. By symmetry, this is true for b and c as well. We conclude that a, b, c have the same set of prime divisors. Let px || a, py || b and pz || c. (Here we write px || a to mean px | a and px+1 | a.) We may assume min{x, y, z } = x. Now b | c2 implies that y ≤ 2z ; c | a2 implies that z ≤ 2x. We obtain y ≤ 2z ≤ 4x. Thus x + y + z ≤ x + 2x + 4x = 7x. Hence the maximum power of p that divides abc is x + y + z ≤ 7x. Since x is the minimum among x, y, z , px divides a, b, c. Hence px divides a + b + c. This implies that p7x divides (a + b + c)7 . Since x + y + z ≤ 7x, it follows that px+y+z divides (a + b + c)7 . This is true of any prime p dividing a, b, c. Hence abc divides (a + b + c)7 . 3. Let a and b be positive real numbers such that a + b = 1. Prove that aa bb + ab ba ≤ 1. Solution: Observe 1 = a + b = aa+b ba+b = aa bb + ba bb . Hence 1 − aa bb − ab ba = aa bb + ba bb − aa bb − ab ba = (aa − ba )(ab − bb ) Now if a ≤ b, then aa ≤ ba and ab ≤ bb . If a ≥ b, then aa ≥ ba and ab ≥ bb . Hence the product is nonnegative for all positive a and b. It follows that aa bb + ab ba ≤ 1. 4. Let X = {1, 2, 3, . . . , 11}. Find the the number of pairs {A, B } such that A ⊆ X , B ⊆ X , A = B and A ∩ B = {4, 5, 7, 8, 9, 10}. Solution: Let A ∪ B = Y , B \ A = M , A \ B = N and X \ Y = L. Then X is the disjoint union of M , N , L and A ∩ B . Now A ∩ B = {4, 5, 7, 8, 9, 10} is ﬁxed. The remaining 5 elements 1, 2, 3, 6, 11 can be distributed in any of the remaining sets M ,

N , L. This can be done in 35 ways. Of these if all the elements are in the set L, then A = B = {4, 5, 7, 8, 9, 10} and this case has to be omitted. Hence the total number of pairs {A, B } such that A ⊆ X , B ⊆ X , A = B and A ∩ B = {4, 5, 7, 8, 9, 10} is 35 − 1. 5. Let ABC be a triangle. Let E be a point on the segment BC such that BE = 2EC . Let F be the mid-point of AC . Let BF intersect AE in Q. Determine BQ/QF . Solution: Let CQ and ET meet AB in S and T respectively. We have [SBC ] BS [SBQ] = = . [ASC ] SA [ASQ] Using componendo by dividendo, we obtain BS [SBC ] − [SBQ] [BQC ] = = . SA [ASC ] − [ASQ] [AQC ] Similarly, We can prove BE [BQA] = , EC [CQA] CF [CQB ] = . FA [AQB ]

But BD = DE = EC implies that BE/EC = 2; CF = F A gives CF/F A = 1. Thus [BQC ] [BQC ]/[AQB ] CF/F A 1 BS = = = = = 2. SA [AQC ] [AQC ]/[AQB ] EC/BE 1/2 Now BQ [BQC ] [BQA] [BQC ] + [BQA] [BQC ] + [BQA] = = = = . QF [F QC ] [F QA] [F QC ] + [F QA] [AQC ]

This gives BQ [BQC ] + [BQA] [BQC ] [BQA] BS BE = = + = + = 2 + 2 = 4. QF [AQC ] [AQC ] [AQC ] SA EC (Note: BS/SA can also be obtained using Ceva’s theorem. One can also obtain the result by coordinate geometry.) 6. Solve the system of equations for positive real numbers: 1 x = + 1, xy z 1 y = + 1, yz x 1 z = + 1. zx y

Solution: The given system reduces to z = x2 y + xyz , x = y z + xyz , y = z 2 x + xyz . Hence z − x2 y = x − y 2 z = y − z 2 x. If x = y , then y 2 z = z 2 x and hence x2 z = z 2 x. This implies that z = x = y . Similarly, x = z implies that x = z = y . Hence if any two of x, y, z are equal, then all are equal. Suppose no two of x, y, z are equal. We may take x is the largest among x, y, z so that x > y and x > z . Here we have two possibilities: y > z and z > y . Suppose x > y > z . Now z − x2 y = x − y 2 z = y − z 2 x shows that y 2 z > z 2 x > x2 y .

But y 2 z > z 2 x and z 2 x > x2 y give y 2 > zx and z 2 > xy . Hence (y 2 )(z 2 ) > (zx)(xy ). 2 3 This gives yz > x . Thus x < xyz = (xz )y < (y 2 )y = y 3 . This forces x < y contradicting x > y . Similarly, we arrive at a contradiction if x > z > y . The only possibility is x = y = z . √ For x = y = z , we get only one equation x2 = 1/2. Since x > 0, x = 1/ 2 = y = z . ———-0000———-